Intel derailed its i9-9900K launch with 2 unforced errors | Industry
The i9-9900K is upon us, along with the rest of Intel’s 9th-gen Core CPUs. The company revealed its Coffee Lake-R processors during an event yesterday that went quite well. Intel revealed its lineup of upcoming Core and Xeon chips. It also spent quite a lot of time talking about how the i9-9900K is the “best gaming CPU ever.” A day later, you would expect that people would want to talk about whether they need to upgrade and whether the price is fair, but that’s not the case. Instead, PC-hardware enthusiasts are furious at Intel for a couple of bonehead moves that it could have avoided.
The two major problems with the event have to do with the i9-9900K. This is Intel’s new high-end gaming chip to compete with AMD’s Ryzen 7 2700X. Like that Ryzen chip, the 9900K has eight cores and 16 threads. It runs at 3.6 GHz, but it turbos up to 5.0 GHz on one-to-two cores. Across all cores, it has a turbo of 4.7 GHz. To me, it sounds a lot like Intel’s 8th-gen i7-8086K with two extra cores. That’s nothing mind-shattering, but I think the 8086K is a great piece of hardware.
I especially like the idea of the 9900K because it’s evidence that Intel is listening to its loudest customers. Most of the 9th-gen Core series is using solder between the CPU dies and their heat spreaders. That should make the chips run significantly cooler than recent generations that relied on a paste-like thermal-interactive material.
So if the product is fine and even exciting in some ways, what’s the problem? Well, Intel is getting in its own way.
During its presentation, Intel said that the Core i9-9900K would sell for $488 each. But what many people, including GamesBeat, didn’t notice was that price came with an asterisk. The $488 is how much Intel is charging per i9-9900K if you buy a thousand of them. I don’t know about you, but I would only ever need one. Even if we were to talk about testing them for work, I wouldn’t require more than two in any reasonable scenario.
So what is the price if you want to buy just one? $580.
You can drive about $90 the gap in those two price points. And I just don’t get why Intel would do this. Yesterday, we wrote that the chip costs $488. Amazon was selling them $499. You could get them for $530 from BHPhoto and Best Buy. And while that’s a lot of money, people were working through their sticker shock. I even came close to writing that the i9-9900K makes the 8086K look like a medium popcorn at the movie theater. Why would I get the medium, when I can spend just $50ish more to get the large with two extra cores and solder?
But it’s only a day later, and now everyone who was considering an i9-9900K has to work through their sticker-shock all over again. And it’s worse because now it feels like an immediate price increase. It is bizarre messaging, and it’s a lousy sales pitch.
Principled Technologies testing
And then we come to the testing that Intel sent out yesterday. The company commissioned a tech firm called Principled Technologies to run third-party benchmarks of the i9-9900K versus other chips including the 8700K and the Ryzen 7 2700X. According to this data, the 9900K is 50 percent faster than the 2700X. And while that seems unlikely for real-world use, that’s not what has the PC hardware community on fire. Instead, enthusiasts are tearing apart Principled Technologies and Intel for presenting some obviously shoddy testing methods as a legitimate point of comparison.
The testing does have a lot of problems. For me, the most shocking thing is that the results use a mix of framerates as reported by the analytics program FRAPS and as reported by in-game benchmark tools. You cannot mix-and-match your measuring devices. That’s bad science that invalidates everything reported.
But Principled Technologies also limited the AMD hardware in some serious ways. The worst example of this is that it was using “Game Mode” for the 2700X, which essentially turns off half of the chip. That means the 2700X was running as a 4-core/8-thread CPU instead of 8 cores with 16 threads.
So yeah, you can throw all of this testing out. It is meaningless.
But here’s the thing. All commissioned benchmarking from a company is meaningless. It’s just marketing. And while marketing works, most people are going to ignore it. The only reason why anyone would end up talking about your marketing a day later is if you did something malicious, dumb, or both.
Why would Intel do this?
Intel should know at this point that the PC-hardware community is looking for these flubs. The company is on top. It charges a premium price compared to the competition. Customers want a reason to hate Intel and to root for AMD. Intel shouldn’t hand those people the narrative that it’s the big bad guy.
What makes even less sense is that Intel doesn’t need to fudge numbers to make its point. It wants to position the i9-9900X as the best gaming CPU? Especially compared to the 2700X? No one who is even a little familiar with that space has any doubt that Intel’s top-of-the-line chip is going to end up outperforming AMD in games. Even when we start getting real benchmarks, it is unlikely that the 2700X will compare favorably in gaming tests.
But Intel’s third-party client did fudge the numbers, and it’s all anyone can talk about — at least when they’re not talking about the price.
I don’t particularly like these narratives. People want the “Intel gonna Intel” story that is easy to follow so they can dismiss the company and feel good about not updating their older hardware or their new AMD purchase. But Intel is ensuring that those people have a legitimate beef. So we end up talking about all of this garbage on the periphery, and we forget about the products themselves.